Hitachi Ltd has been celebrating its four score years and 10 with the Hitachi Technology 1991 exhibition in New York, London and Tokyo. The exhibitions featured an impressive array of Hitachi’s current and future technologies, and the company unleashed its more formidable scientists on an audience that had little option but to believe them. The first seminar was presented by Yasutsugu Takeda, executive managing director of Hitachi Ltd, and he provided an overview of the $4,400m company and its 740 subsidiaries. The largest business division is information systems and electronic devices, which constitutes 33% of the total. Research and development expenditure in 1989 was $2,200m, around 9.7% of annual sales. Takeda refused to be drawn on what he regards the ideal level of investment in research, saying that it ought to be determined by return on investment and cash flow.
Quantum electronics
Information systems and electronics consumes 70.6% of research expenditure, and the company has established an Advanced Research Laboratory to focus on areas where it has little or no presence. Hitachi has a policy of setting up branch research facilities and to date, it has one in Cambridge specialising in quantum electronics, Dublin has information science, and labs in Dusseldorf and Milan focus on industrial applications. There are facilities in San Francisco and Detroit, specialising in semiconductors and automotive components respectively, and Japan has a further nine research laboratories. Mr Takeda was followed by his colleague Koichiro Ishihara, chief researcher at Hitachi’s Systems Development laboratory. He believes that the information market will grow at 10% per year, and by 2000, will have reached the $700m mark. He says that the fastest growing segment, software and services, will fuel demand for higher performance and larger storage capacities, and his colleague, Yoshito Tsunoda, department manager at the Central Research Library, picked up on this theme. He forecasts that the worldwide market for storage products will reach $70,000m by the turn of the century, and although optical disks will capture a large share, magnetic disks will remain the medium of choice. Tsunoda reckons that the 21st century will usher in larger capacities and higher data throughput on small form factor hard disk drives, the larger-scale hard disk drive will be replaced by array disk systems, and optical disk drives will also achieve much larger capacities. The area density of magnetic storage in 1990 is around 100M-bits per square inch, but by 2010, that will have increased to 10G-bits, and the same applies to optical technology. Tsunoda says that there are various approaches to achieving area densities of more than 1G-bits per square inch, one of which is higher track pitch recording with a more sensitive head. Others include the development of a narrow track width head and vertical recording with less spacing between head and disk. Since area density of optical storage is determined by spot size, higher area density can be obtained by decreasing the focused size – in other words, shortening the wavelength of the light source. He says that intensive attempts to develop shorter wavelength light sources are being made, including Second Generation Harmonic optics. –
By Janice McGinn
A beam with a 0.8 micron wavelength from a laser diode is focused on a Nd-YAG crystal and this can be converted to a 0.53 micron wavelength by the non-linear effect of the crystal, achieving an areal storage density of 3G-bits per square inch on the platter. Yoshizumi Eto, chief researcher at Hitachi’s Central Research Laboratory, went on to talk of his visions for the future of video technologies and high definition television. Eto says that HDTV is an entirely new system aiming for top quality rather than compatibility with existing systems, although he also stresses that Hitachi’s products are designed to be widely accepted in a world where various television systems both exist and are emerging. Which is an oblique reference to the disagreement between Japan
and the US and European developers of HDTV. Europe and the US want upwards-compatible systems, the Japanese say that ought to be secondary to quality, although Eto acknowledges that the problem could, and probably will, be resolved by using encoder-decoders. The final seminar was presented by Shojiro Asai, deputy general manager of the Central Research Laboratory, and he talked about the courseHitachi has set for its research activities in microelectronics. In the area of microlithography, the company is aiming at 0.1 micron, 1G-bit memory chips with a processing capability of 1 GIPS and transmission rates above 40Gbps. At present, the 1M-bit DRAM is the most widely used in the industry, with the 4M-bit DRAM is mass production, the 16M-bit version in development, and the 64M-bit laboratory prototype announced by Hitachi earlier this year. Eto expects 64M-bit DRAMs to be in production by 1994 or 1995, and he says that Hitachi is already looking at a 256M-bit offering. To obtain higher speeds and performance from semiconductors, it is essential to have a means of engraving small features on them, and one of these techniques is phase-shift optical lithography. Phase shift enables the printing of mask patterns onto Silicon as close together as 0.17 micron, about 70% of the wavelengths of the light source. Eto says that means 256M-bit or 1G-bit DRAMs may be manufactured using optical projection lithography, although he acknowledges that phase shifting is not really suitable for irregular patterns. Hitachi is also looking at electron-beam lithography, and its E-beam machine can reduce dimensions to 0.1 micron.
Rejected X-ray
Eto denies that Hitachi has rejected X-ray lithography in favour of optical techniques – unlike IBM – and says that while he talked of optical lithography’s limitations some 10 years ago, he believes that it can be improved still further. Apart from shrinking transistors and capacitors, Eto says that it is also necessary to reduce the thickness of thin dielectric films in proportion to device dimensions, and he claims that Tantalum Pento-oxide will become indispensable for 64M-bit DRAMs and beyond. New devices and mechanisms under development include optical switches consisting of sets of parallel optical waveguides built onto chips. Input optical signals are steered by controlling the injection current at intersections onto the output waveguides. Experiments are under way to see if it is possible to build an optical switching system for B-ISDN in asynchronous transfer mode, and these devices could be employed in optical computing. Finally, another area under exploration is superconducting electronics, and Eto says that Hitachi now has a 4-bit digital signal processor chip operating at 1Gflops at low liquid Helium temperature.
